Multiple types of oncolytic viruses are currently under investigation in clinical trials. To optimize therapeutic outcomes it is believed that the plethora of different tumor types will require a diversity of different virus types. Sendai virus (SeV), a murine parainfluenza virus, displays a broad host range, enters cells within minutes and already has been applied safely as a gene transfer vector in gene therapy patients. However, SeV spreading naturally is abrogated in human cells due to a lack of virus activating proteases. To enable oncolytic applications of SeV we here engineered a set of novel recombinant vectors by a two-step approach: (i) introduction of an ubiquitously recognized cleavage-motive into SeV fusion protein now enabling continuous spreading in human tissues, and (ii) profound attenuation of these rSeV by the knockout of viral immune modulating accessory proteins. When employing human hepatoma cell lines, newly generated SeV variants now reached high titers and induced a profound tumor cell lysis. In contrast, virus release from untransformed human fibroblasts or primary human hepatocytes was found to be reduced by about three log steps in a time course experiment which enables the cumulation of kinetic differences of the distinct phases of viral replication such as primary target cell infection, target cell replication, and progeny virus particle release. In a hepatoma xenograft animal model we found a tumor-specific spreading of our novel recombinant SeV vectors without evidence of biodistribution into non-malignant tissues. In conclusion, we successfully developed novel tumor-selective oncolytic rSeV vectors, constituting a new tool for virotherapy of solid tumors being ready for further preclinical and clinical development to address distinct tumor types.